Light-emitting diodes (LEDs) are gradually replacing incandescent light bulbs in various applications, including traffic signal lamps, large-sized full-color outdoor displays, various lamps for automobiles, solid-state lighting devices, flat panel displays, and the like. Conventional LEDs typically include a light-emitting semiconductor material, also known as the bare die, and numerous additional components designed for improving the performance of the LED. These components may include a light-reflecting cup mounted below the bare die, a transparent encapsulation (typically silicone) surrounding and protecting the bare die and the light reflecting cup, and electrical leads for supplying the electrical current to the bare die. The bare die and the additional components are efficiently packed in an LED package.
LEDs also represent an attractive alternative light source for general lighting applications and for backlights for liquid crystal displays, where they enable extremely low-thickness (or “low-profile”) solutions. One conventional geometry for such illumination solutions is the so-called edge-lit configuration, in which a packaged LED is attached to the shorter, narrow side (or “face”) of a waveguide, and the light is emitted through the broader “top” face of the waveguide. Increased coupling efficiencies may be obtained by embedding the bare LED die within the waveguide itself, rather than by separately encapsulating or packaging the die before coupling it to the waveguide. However, since the geometric dimensions of a waveguide typically are far larger than those of the LED die, it is often challenging to achieve and maintain the high coupling efficiency enabled by embedding the bare die while also forming a strong mechanical connection between the various components of the completed system. The LED die is typically mounted on a platform, or a “sub-assembly” that provides mechanical support and electrical connectivity to an external power source. The presence and geometry of the LED sub-assembly may present difficulties when attempting to embed the LED die within the waveguide with high coupling efficiency.
FIGS. 1A (cross-section) and 1B (bottom view) depict an illumination device 100 that features an LED 105 mounted on a sub-assembly 110 and coupled within a recess 115 in a waveguide 120. As shown, the waveguide 120 has the shape of a thin plate with flat top and bottom faces, which may be parallel, as shown, or may be angled toward each other, giving the waveguide 120 the shape of a wedge. Mirrors (or mirror coatings) may be present along the bottom and side faces of waveguide 120. During operation of the illumination device 100, light from the LED 105 is coupled in to an input region 130 of the waveguide 120 via an input coupling element 135. The light then propagates toward an output region 140 by means of total internal reflection (TIR) off of the top and bottom faces of the waveguide 120. (As known to those of skill in the art, TIR depends at least on the refractive-index difference of two materials at the boundary therebetween, as well as the angle of the light impinging upon the boundary.) In the output region 140, the light is out-coupled from the waveguide 120 by, e.g., embedded scattering elements 145 that disrupt the TIR propagation, resulting in emitted light 150.
Devices such as illumination device 100 present an extremely difficult challenge—the need to, within the small thickness of the waveguide 120, convert light not emitted from the LED 105 in the TIR condition into light propagating in waveguide 120 via TIR. This conversion generally must be performed within a very small area in order to prevent additional loss of light from impingements of light on the waveguide faces in non-TIR conditions. This, in turn, constrains the area of the LED sub-assembly 110, as the top face of the sub-assembly 110 generally does not reflect light at TIR conditions and/or may even absorb light from the LED 105, diminishing overall efficiency.
However, for devices such as illumination device 100 to have adequate mechanical stability, the area of the “joint” between the waveguide 120 and the LED sub-assembly 110 is typically much larger than that of the LED die itself, resulting in the above-described efficiency-diminishing area of the sub-assembly surrounding the LED die. This additional sub-assembly area increases the cross-section of contact between the waveguide and the sub-assembly, strengthening the connection, but also results in decreased input coupling efficiency.
FIG. 2 depicts one conventional approach to addressing this trade-off, in which only the “optical connection” (i.e., the proximity enabling in-coupling of light from the LED) between the LED and the waveguide is made in immediate proximity to the LED, and the mechanical support between the waveguide and the sub-assembly is provided separately. As shown, an illumination device 200 features an LED 210 mounted on a larger sub-assembly 220 that is joined to a waveguide 230. The LED 210 is flanked by two mechanical connections 240 that provide mechanical support when the sub-assembly 220 is joined to the waveguide 230. In many such designs, it is recognized that light emitted from the LED in one or more lateral directions (e.g., the indicated x-direction) will not reach the output region even in the absence of mechanical connections that may block or absorb such light. Specifically, much of the light reaching the side faces of the waveguide in such directions will not reach the output region due to their multiple reflections, in the x-direction, in the input-region. Thus, conventional designs may place the mechanical connections 240 in such locations, as light lost via interaction therewith may well not have been efficiently coupled into the bulk of the waveguide anyway; thus, losses associated with the mechanical connections may have little additional impact on the input coupling efficiency.
Exacerbating the impact of mechanical supports on the input coupling efficiency is the fact that the optical connection between the LED and the waveguide is typically achieved via a “dam and fill” process, in which a low-viscosity dam of encapsulant is formed between (and in contact with) the sub-assembly and the waveguide and then filled with higher-viscosity index-matching material. Such processes fill the entire in-coupling region near the LED with the index-matching material, which increases the size of the surrounding region incapable of TIR-based confinement of the LED light (because, since the index-matching material contacts the waveguide and the light-absorbing sub-assembly, light propagating toward the sub-assembly is simply absorbed or otherwise lost rather than reflected into a TIR mode and efficiently in-coupled). Thus, in view of the challenges and disadvantages of conventional waveguide-based illumination devices described above, there is a need for illumination devices having increased mechanical stability without associated in-coupling losses that adversely impact overall efficiency.